All posts by medical

Anatomy

The world’s worst haka: anatomy of the 1973 All Blacks fiasco – Telegraph.co.uk

The Haka has become one of the most beloved and inspirational spectacles in world sport.

New Zealanders talk passionately about its spiritual significance, the way it links them to their past and their ancestors, and its central place in their sporting culture.

Here, we examine how the Haka grew from this 1973 abomination against the Barbarians at Cardiff Arms Park.

Watch the video below, and then well examine it frame-by-frame, in excruciating detail.

Original post:
The world's worst haka: anatomy of the 1973 All Blacks fiasco - Telegraph.co.uk

Human Behavior

DeepMind Asks: How Much Can Humans Teach AI? – Futurism

In BriefDeepMind is collaborating with humans so that its AI can learnusing human feedback instead of collecting rewards as it exploresits environment. This work will help AI systems perform moreeffectively and safely, and do what we want them to do. Humans Teaching Robots Artificial Intelligence (AI) has the potential to advance humanity and civilization than any technology that came before it. However, AI carries risks, and heavy responsibilities, with it. DeepMind, owned by Alphabet (Googles parent company), and OpenAI, a non-profit AI research company, are working to alleviate some of these concerns. They are collaborating with people (who dont necessarily have any special technical skills themselves) touse human feedback to teach AI. Not only because this feedback helps AIlearn more effectively, but also because the method providesimproved technical safety and control.

Not only because this feedback helps AIlearn more effectively, but also because the method providesimproved technical safety and control.

Among the first collaboration conclusions: AI learns by trial and error, and doesnt need humans to give it an end goal. This is good, because we already know that setting a goal thats even a little off can have disastrous results. In practice, the system used feedback to learn how to make a simulated robot do backflips.

The system is unusual because it learns by training the reward predictor, an agent from a neural network, instead of collecting rewards as it explores an environment. A reinforcement learning agent still explores the environment, but the difference is that clips of its behavior are then sent to a human periodically. That human then chooses the better behavior based on whatever the ultimate goal is. Its those human selectionsthat train the reward predictor, who in turn trains the learning agent. Finally, thelearning agent eventually learns how to improve its behavior enough to maximize its rewards which it can only do by pleasing the human.

This approach allows humans to detect and correct any behaviors that are undesirable, which ensures safety without being too burdensome for human stewards. Thats a good thing, because they need toreview about 0.1% of the agents behavior to teach it. That may not seem like much at first, butthat could well mean thousands of clips to review something the researchers are working on.

Human feedback can also help AI achieve superhuman results at least in some video games. Researchers are now parsing out why the human feedback system achieves wildly successful results with some tasks, average or even ineffective results with others.For example, no amount of human feedback could help the system master Breakout or Qbert. They are also working to fix the problem of reward hacking, in which early discontinuation of human feedback causes the system to game its reward function for bad results.

Understanding these problems is essential to building AI systems that behave as we intend them to safely and effectively. Other future goals may include reducing the amount of human feedback required, or changing the way its provided; perhaps eventually facilitatingface to face exchanges that offer the AI more opportunities to learn from actual human behavior.

Editors Note: This article has been updated to note the contributions made by OpenAI.

See the original post:
DeepMind Asks: How Much Can Humans Teach AI? - Futurism

Genetics

NewLink Genetics Is Still Undervalued, Despite The Disappointment From Navoximod – Seeking Alpha

NewLink Genetics (NASDAQ:NLNK) lost 40% of its value in a single day, as it surprisingly announced that Roche ([[OTCQX:RHHBY]], [[OTCQX:RHHBF]])/Genentech would return the rights on Navoximod, an IDO inhibitor, to the company. In this article, I will discuss what the reasons for this decision were and importantly why I believe that NewLink Genetics is highly undervalued at the current, all-time low, share price of $6.46 (Figure 1).

Figure 1. Common stock chart for NLNK. Source: Yahoo Finance

NewLink Genetics pipeline concentrates on the development of IDO (indoleamine-2, 3-dioxygenase) pathway inhibitors. IDO inhibitors, in general, are expected to boost the body's immune system to fight against cancer, similar to PD-1 and CTLA-4 pathway inhibitors.

NewLink was focused on two distinct IDO pathway inhibitors:

One is the already mentioned Navoximod, a direct inhibitor of IDO, which works very similar to Incyte's (NASDAQ: INCY) Epacadostat and Bristol-Myers Squibb's (NYSE: BMY) BMS-986205. Their most advanced drug is Indoximod, which is no direct blocker of IDO but rather mimics the effect of IDO inhibition.

Navoximod

In 2016 Roche/Genentech licensed all worldwide commercialization rights for Navoximod and paid NewLink Genetics $150 million upfront, with eligibility to $1 billion more if certain milestones are met.

Navoximod is currently tested in phase 1 trials in several solid cancers including non-small-cell lung cancer ("NSCLC"), renal cell cancer ("RCC"), urothelial bladder cancer ("UBC"), triple-negative breast cancer ("TNBC") in combination with Atezolizumab (the PD-L1 inhibitor from Genentech/Roche). In all these cohorts combined (separated data is not available yet) Navoximod + Atezolizumab showed a partial response in only 9% (4/45) of patients [1].

Epacadostat is tested in many different solid cancers in partnership with Merck's (NYSE: MRK) Keytruda. If the same patient cohorts of Epacadostat and Keytruda are pooled they achieve 27% (40/146 pooled; stratified in cancer types: 14/40 NSLC, 13/37 UBC, 9/30 RCC, 4/39 TNBC) [2][3]. Based on this data Roche/Genentech decided to return the rights for Navoximod to NewLink Genetics.

I think one of the reasons why Navoximod doesn't perform as well as Epacadostat could be rooted in their differential efficacies to inhibit IDO. This is read out by measuring the kynurenine levels in the blood (kynurenine is the product of an active IDO enzyme, so the greater the drop in kynurenine levels the better inhibited is IDO). Navoximod brings down blood kynurenine to 70% of the pre-treatment level. In comparison to that, Epacadostat achieves a 50% reduction and even better is BMS-986205, which manages to drop the kynurenine level to about 40% of pre-treatment levels [4][5]. This indicates that Navoximod is not such a potent IDO inhibitor than the ones of the competitors and thus might explain why Navoximod has less efficacy than Epacadostat.

Based on the disappointing preliminary results and especially with a competitor that is, first, much further in the development (multiple clinical phase 3 trials of Epacadostat + Keytruda will be started in 2017) and second, shows better efficacy, it is understandable that Roche/Genentech decided to not further develop Navoximod. I personally think it is unlikely that NewLink will continue to develop Navoximod unless they see potential in any of the single cancer types. My feeling is that in neither of the single cancer types Navoximod performed comparable to Epacadostat, otherwise Roche/Genentech wouldn't have returned the rights to the drug.

Indoximod

Indoximod is tested in combination with several different agents in phase 2 studies in melanoma, pancreatic cancer, breast cancer, metastatic castration-resistant prostate cancer (mCRPC) and glioblastoma.

Melanoma

Beginning of April 2017, NLNK released preliminary results of Indoximod in combination with Keytruda in advanced melanoma. The combination achieved a 59% objective response rate and 80% disease control rate (Table 1) [6]. Including patients with ocular melanoma, a very hard to treat patient population, the ORR is 52% and the DCR is 73%.

Treatment

Objective response rate

Disease control rate

Grade 3 adverse events

Epacadostat / Keytruda

n = 19

58%

74%

19%

Indoximod / Keytruda

n = 51

59%

80%

Nivolumab / Ipilimumab

n = 314

58%

71%

55%

Table 1. Objective response rates and adverse events in melanoma. For Indoximod/Keytruda only data from non-ocular patients are included. *Only incomplete data available

This news was followed by a 33% drop in NLNK's stock price, due to disappointment that the Indoximod/Keytruda combination was not much better than the 58% ORR and 74% DCR achieved with Epacadostat/Keytruda in treatment-naive advanced melanoma [7]. Another reason for the disappointment is the complete response rate of 26% for Epacadostat/Keytruda, compared to only 12% for Indoximod/Keytruda. Here it should be taken into account, however, that the presented data for Epacadostat consist of only 19 patients, whereas the data for Indoximod include 51 patients. So it is possible that the complete response rates of Epacadostat will drop once more patients are added.

I was positively surprised by how similar the two drugs performed, given that the mechanism of action differ quite significantly between Indoximod and Epacadostat and would rate it as a good sign that in melanoma Indoximod is able to perform similarly than Epacadostat.

The current, best-in-class, treatment option for metastatic melanoma is the combination of Nivolumab with Ipilimumab (PD-1 and CTLA-4 inhibitors from BMY), which achieves 58% ORR and 71% DCR (Table 1) [8].

The phase 3 trials of Indoximod and Epacadostat in melanoma will likely be benchmarked against this combination. The IDO inhibitor combinations seem to be unable to surpass anti-PD-1/CTLA-4 inhibition in terms of ORR or DCR. But then the rate of serious adverse events will come into play. Epacadostat/Keytruda has 19% of Grade 3 adverse events and although no concrete numbers are available for Indoximod/Keytruda, they are not higher than with Keytruda alone.

This means that both IDO inhibitors are exceptionally well tolerated and well below the 55% of Grade 3 adverse events observed with the anti-PD-1/CTLA-4 treatment. Therefore, there is a good chance that the IDO inhibitors will be approved based on the more manageable safety profile.

Acute myeloid leukemia

In acute myeloid leukemia, Indoximod is tested in combination with 7+3 chemotherapy. This combination leads to a complete remission in 83% of tumors (5 of 6 patients) with no evidence of minimal residual disease [9]. With 7+3 chemotherapy alone in young adults, complete remission can also be achieved in up to 75% of patients [10]. The big problem rather is that the relapse rate is very high and thus it will be critical for Indoximod to show prolonged relapse-free survival and overall survival.

Pancreatic cancer

In pancreatic cancer, the combination of Indoximod with chemotherapy achieved an ORR of 45% (14/31) vs. 23% for chemotherapy alone [11]. Pancreatic cancer is a notoriously hard to treat cancer type and many drugs failed to get approved. NewLink, for instance, also tried to get a drug approved for pancreatic cancer, called Algenpantucel-L. After promising objective response rates in a phase 2 study, the drug failed to enhance overall survival in a phase 3 trial. So I think that also for Indoximod it is important to remain patient and to wait whether the convincing ORR in pancreatic cancer can be transformed into a survival benefit.

Metastatic castration-resistant prostate cancer

Indoximod is combined with PROVENGE (a vaccine already approved for prostate cancer) for patients with mCRPC. The combination was able to enhance radiographic progression-free survival (rPFS) from 4.1 months in the placebo arm to 10.3 months in the treatment arm [12]. This compares to the rPFS of Enzalutamide, an androgen receptor inhibitor, which is 8.3 months [13]. In general, rPFS is highly associated with overall survival in mCRPC and so it is likely that Indoximod/PROVENGE will also enhance overall survival [14].

Glioblastoma

In addition to these trials, Indoximod is also tested in glioblastoma also in combination with chemotherapy and a 6-month progression free survival in 25% of patients compared to a historical rate of 15% [15].

Breast cancer

Recently a statement was released, that Indoximod plus chemotherapy failed to meet the primary end points of statistically different progression-free survival and overall survival in metastatic breast cancer [16].

The use of IDO inhibitors in breast cancer thus continues to disappoint, as also the combination of Epacadostat and Keytruda only achieved a 10% objective response rate in triple negative breast cancer [17].

Valuation and Conclusion

NewLink Genetics has a current market capitalization of $ 189 million. If cash of $ 75 million, expected at the end of 2017 and debt plus royalty obligations of $ 6.5 million are taking into account, the whole company is currently valued at $ 120.5 million.

Their most promising and also furthest developed indication for Indoximod is in melanoma. I think it was very important for NewLink to show comparable objective response rates, which, since they have a different mechanism of action, was far from certain. As estimated by Bhavneesh Sharma in his article, the current, risk-adjusted, peak revenue of Indoximod in melanoma alone is $ 106 million in 2024. This means the company is currently valued at little above the peak revenue of a single indication. They furthermore have promising first results in other indications as well, importantly in cancer types that Incyte is currently not pursuing or lagging behind.

Given the general high valuation of companies developing novel immunotherapies, NewLink Genetics market capitalization is extremely low. This was majorly caused by disappointment that Indoximod didn't surpass Epacadostat's objective response and complete response rates in melanoma and of course by Roche/Genentech returning their rights on Navoximod to the company.

At this price, NewLink Genetics to me seems to be an attractive takeover candidate for companies who want to quickly spice up their immune-oncology portfolio but are not willing to pay a high premium for it. Bristol-Myers Squibb, for instance, paid $ 800 million upfront (total deal volume is $ 1.25 billion) for Flexus Therapeutics, which developed a preclinical IDO inhibitor of in 2015 [18].

If NewLink Genetics is not bought up or they continue to develop Indoximod unpartnered, they will need to raise additional capital at some point until the end of 2018. Other risks include the failure of clinical trials or that their drugs do not get approved by the regulatory agencies.

Summing this up, I believe the current valuation of NewLink Genetics is very cheap and I fully expect their stock price to rise again, once investors have digested the disappointments of the last months and regain their faith in the company again.

Disclosure: I am/we are long NLNK.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

See more here:
NewLink Genetics Is Still Undervalued, Despite The Disappointment From Navoximod - Seeking Alpha

Genetics

At-home genetic test reveals secrets about your health – FOX 5 Atlanta

ATLANTA - You could call 23andMe a test for the genetically-curious.

For $199, you can find out what your genes reveal about not just your ancestry and your traits, but your health.

Dr. Jeffrey Pollard, Director of Medical Affairs for 23andMe says 2 million people have already taken the test.

"I think it's exciting to learn something about yourself," Dr. Pollard says. "Everyone likes learning new things about themselves."

The process is pretty simple.

You order the kit, collect a sample of your saliva and send it off to 23andMe.

A few weeks later, through an online portal, you can read you genetic health risk test results, and find out if you might be at higher risk of developing certain diseases like late-onset Alzheimer's Disease and Parkinson's Disease.

But genetics testing in relatively new, says Pollard, and there are limits to what 23andMe can reveal.

"We are not a diagnostic test," Pollard cautions. "So, we aren't telling anyone the have a certain condition. We're really presenting them with an element of their genetic risk, if you will."

Kimberly King-Spohn, Director of the WellStar Center for Genetics, says at least a third of her patients have mentioned 23andMe since April, when the US Food and Drug Administration cleared the direct-to-consumer genetic testing company to offer health risk information to consumers for 10 disease and conditions.

King-Spohn has some concerns.

"It's confusing," she says. "They get a very long report with a lot of information, and patients have had a difficult time interpreting what that means for their health."

Dr. Pollard says 23andMe provides a lot of context, to help customers make sense of their findings.

But, King-Spohn argues patients may not want to know they're at higher genetic risk of developing diseases they can't really do anything about.

"We don't have a treatment or intervention for Alzheimer's disease," she says. "So, what would you do with that information?"

Pollard says the information 23andMe offers is just a piece of a much larger puzzle.

"We like to highlight the fact that your genetic risk is just one element in play," he says.

Because, Pollard says, there are a lot of factors you can control, when it comes to your health.

"You might be able to change certain other things that you do in your life, whether it's exercise, or your activity level, your diet, whether or not you smoke," he says.

Both Pollard and King-Spohn agree it's important to read the fine print, before you get screened, to make sure you understand what 23andMe can and cannot tell you.

It's a lot to absorb, they both say.

"But it's also empowering" Dr. Pollard says. "And, we hear that our customers are taking this information and doing something with it in their lives. So, that, I think, is a very powerful beginning."

See the rest here:
At-home genetic test reveals secrets about your health - FOX 5 Atlanta

Cell Biology

Watch This 3D Cell Culture Space – Genetic Engineering & Biotechnology News (press release)

GEN: Are 3D cell culture models as strongly focused as ever on drug safety testing, or are they finding new applications?

Dr. Aho: The focus of 3D cell culture models has definitely expanded beyond drug safety testing. It is becoming increasingly clear that these models mimic cells in vivo at a greater capacity than traditional cell culture.

In addition to drug toxicity, 3D models are progressively being developed and used in developmental biology research, disease modeling, and regenerative medicine. 3D models also provide an enhanced system for drug discovery. Because they better recapitulate disease in vitro, 3D models have the potential to accelerate the testing timeline for drug efficacy studies.

Dr. Banks: Another major application area of 3D cell culture models is in oncology. Spheroids in both media and Matrigel can be used as surrogate models of tumor proliferation and tumor invasion. Automated brightfield or fluorescence microscopy is typically used for spheroid or invadipodia area measurements. In addition to spheroids, collagen-based scaffolds that encourage cell aggregation into tumoroids have been used for immunotherapy applications such as natural killer cell cytotoxicity assays. Finally, magnetic particles have been used to bioprint cells for cell migration and invasion experiments.

Dr. Eglen: We would argue that 3D cell culture models have been used for many years in basic research and disease modeling, notably in cancer researchthis was, after all, one of the original applications of Corning Matrigel, a naturally occurring extracellular matrix for us in 3D cell culture. That said, it is true that 3D cell culture models are increasingly being used in preclinical lead optimization, particularly in evaluating potential compound toxicity and metabolic liability.

Furthermore, disease research areas are expanding to include neurology, stem cell research, cell therapy, and (potentially) tissue engineering. Perhaps the most exciting work is the development of 3D technologies for the optimal production of patient-specific cells, either for compound testing or possibly cell therapy.

Interestingly, spheroids derived from stem cells grown in 3D models show improved stemness, that is, characteristics that may lead to increased efficacy in regenerative medicine. Researchers have seen that spheroids display enhanced anti-inflammatory, tissue regenerative, and reparative responses, as well as better post-transplant survival of mesenchymal stem cells.

Autologous tissue for transplantation may also come from organoids produced via 3D cell culture. For example, renal organoids derived from pluripotent stem cells have been successfully transplanted under the renal capsules of adult mice. Clearly, research in this area is advancing rapidly, probably due to a convergence of several multidisciplinary fields, ranging from bioengineering, materials science, phenotypic screening, and cell biology.

Dr. Trezise: Drug safety continues to be a significant application area for 3D models. This application area has become only more interesting as more data has become available indicating that 3D models offer translational benefits. In addition, there is a growing trend to develop 3D models that can advance developmental biology, target validation, and drug efficacy studies. This trend is particularly evident in the field of oncology, where researchers are combining patient-specific tumor cells and 3D cell culture methods to create tumor organoids. These mini-tumors are being used to determine sensitivity to combinations of different chemical, biological, and cellular therapeutics in the context of personalized medicine.

Dr. Klette: 3D cell culture models are widely used for drug safety testing, such as studying hepatic injury from compound screens, and for examining drug metabolism using 3D hepatocyte models. In personalized medicine, however, patient-derived primary 3D models are being used for cancer screening in biotherapeutics. Here, 3D models provide enhanced physiological relevance to determine drug efficacy and potential impacts on carcinogenesis, metastasis, and tumor reoccurrence. If we look outside drug discovery and biologics, we notice that areas such as regenerative medicine and cell therapies can take advantage of 3D models as a predictor of disease and (when scaled to therapeutic levels) as a disease treatment.

Dr. Guye: 3D cell models are applied throughout the biomedical and life sciences. 3D technologies that are compatible with high-throughput screening are used not only for screening purposes, but also for target and hit validation, lead optimization, and investigational toxicology.

Basically, given their ability to extend cell lifetimes and incorporate multiple cell types, 3D models are increasingly finding their way into basic research, where they are helping to recapitulate disease progression and assess the impact of certain genes and pathways on disease progression/preventionactivities that help scientists define adverse outcome pathways. Importantly, we expect human 3D cell culture models to significantly reduce the percentage of drugs that progress to clinical trials and fail due to lack of efficacy.

Dr. Kugelmeier: The focus on drug safety testing is still valid, and sophisticated organoid models might contribute to even more accurate drug safety testing because of increased physiological fidelity of these models. But there are also significant new research areas. Combining organoid technology with stem cell biology could lead to therapeutic applications. Also, cancer researchespecially cancer research that focuses on cancer stem cellsneeds 3D models. Of these models, cell spheroids are among the most important. Sophisticated cell-spheroid platforms not only allow research but also provide drug-testing possibilities using patient cells for personalized medicine. Finally, these platforms may enable therapeutic applications with stem cell spheroids in regenerative medicine.

Mrs. Hussain: The focus for 3D cell culture methods is still the drug safety testing that occurs before in vivo testing. Recently, there has been a renewed interest in phenotypic drug screening to discover new drug targets. With this shift, there is growing emphasis on bridging the gap between phenotypic screens and 3D methods. Phenotypic screens, in vitro, were traditionally carried out using 2D methods that do not take into account the complexity of the in vivo environment. 3D methods are now sought to build biologically relevant models that are more predictive of phenotypic response to new drug targets.

Dr. Bulpin: Applications continue to expand for 3D models, including the development of specific disease models and complex tissue models that can be used for basic research as well as drug discovery. Another promising area for 3D models is personalized medicine. Several types of cells can be used in these models including immortalized cells, genetically engineered cells, induced pluripotent stem cellderived cells, primary human cells, and patient-derived cells (including patient-derived xenografts). Another potential research avenue is engineering 3D tissues for organ transplants.

Dr. Joore: Over the last year, we observed a growing interest in 3D tissue models that could be used in studies of disease processes, whether the studies emphasized screening or efficacy analysis. These are, I think, two sides of the same coin. Once researchers realize they need better predictive models for safety testing, they start to see that improved models would also have potential for discovery and development. Molecule-to-molecule screens have generated lots of very specific inhibitors, but not so many therapies. Researchers are now starting to appreciate the richness of 3D model data, especially in combination with the throughput of our organ-on-a-chip platforms.

Ms. Floyd: Cancer researchers and developmental biologists have certainly benefitted from 3D cell culture models, which are more physiologically relevant than are 2D systems to the study of cellular differentiation. Further, 3D in vitro systems are well positioned to obtain approvals from authorities such as the Organization for Economic Co-operation and Development Organization (OECD). The OECD and other bodies are considering alternatives to whole-animal testing, including alternatives that can accomplish skin-sensitization studies for the safety assessment of chemicals.

Prof. Przyborski: What has changed more recently is the ease of access to innovative technologies on the market that enable researchers to more readily practice 3D cell culture routinely. 3D cell culture has had impact in multiple areas in basic research, drug screening, and safety assessment. Researchers are now looking to 3D technologies to create more sophisticated models that are representative of real human tissues. Investment in more advanced in vitro assays at an early stage will improve predictions of drug action and inform the decision-making process as to whether to further invest in a particular drug candidate.

Dr. Kennedy: 3D cell cultures continue to be extensively explored for drug safety screening; however, there is a growing interest in expanding the use of more complex 3D models into areas such as disease modeling and precision medicine. For example, preclinical hepatic research is now looking to exploit the benefits of spheroid cultures by building 3D co-culture models that consist of multiple primary liver cell types to create new models of hepatic and biliary disease. Likewise, stem cellderived organoids are opening the possibility of tailoring therapeutic regimens to patients genetic makeups and to identify the best treatment options.

Read the original here:
Watch This 3D Cell Culture Space - Genetic Engineering & Biotechnology News (press release)

Biochemistry

Faculty member and his wife give $1 million to UCLA – UCLA Newsroom

Penny Jennings/UCLA

Michael Jung

Michael Jung, a distinguished professor of chemistry and biochemistry in the UCLA College, and his wife, Alice, have donated $1 million toward the establishment of the Michael and Alice Jung Endowed Chair in Medicinal Chemistry and Drug Discovery.

The gift was matched by the UCLA division of physical sciences for a total contribution of $2 million. The match was made possible by a program established after UCLA sold its royalty interest in Xtandi, a compound developed by Jung and his research team for the treatment of prostate cancer. With its share of the proceeds from the Xtandi transaction, UCLA has also made matching funds available for gifts that support undergraduate scholarships at UCLA.

Xtandi has not only saved lives; it has been a wonderful boost to UCLA due to the matching program, and we have Mike Jung to thank for that, said Miguel Garca-Garibay, dean of physical sciences. He and Alice have set a terrific example by endowing a chair in Mikes department, for which we are very grateful.

Jung is an authority on synthetic organic and medicinal chemistry. He is an inventor on 34 issued patents and 36 patent applications arising from both his consulting activities and his own research. He has more than 15 ongoing academic research collaborations and consults for more than 20 industrial laboratories in both biotech and pharmaceutical settings.

His current research holds promise for the development of new drugs for the treatment of various diseases and conditions, including for breast, lung and prostate cancer; antiviral diseases; muscular dystrophy; multiple sclerosis; osteoporosis; and even hair loss.

My wife and I hope that our gift will enable UCLA to hire a faculty member who could continue to do similar drug discovery research well into the future, with the hope of producing more useful drugs, Jung said.

A member of the UCLA faculty since 1974, Jung has published more than 345 research papers and presented more than 600 lectures on his research. He has supervised 92 doctoral and nine masters theses, and he has taught more than 130 postdoctoral scholars.

Among the awards he has received are the American Chemical Societys Arthur C. Cope Scholar Award, UCLAs Glenn T. Seaborg Medal and Gold Shield Faculty Prize, and the 2015 Team Science Award from the American Association for Cancer Research. He also was elected to the National Academy of Inventors.

Without chemistry, we wouldnt have life-saving medicines like Xtandi, said Catherine Clarke, chair of the chemistry and biochemistry department and a professor of biochemistry. Thanks to Mike and Alice Jungs gift, the department will be able to pursue more breakthrough research in medicinal chemistry. Who knows how many more lives will be saved?

The department of chemistry and biochemistry was named No. 7 in the world in chemistry in the 2017 U.S. News and World Report Best Global Universities rankings, and three faculty members and four alumni have been awarded the Nobel Prizes in chemistry. The department has more than 50 faculty, 130 postdoctoral researchers, 350 graduate students and 1,400 undergraduates.

The gift is part of the $4.2 billion UCLA Centennial Campaign, which is scheduled to conclude in December 2019 during UCLAs 100th anniversary year.

View post:
Faculty member and his wife give $1 million to UCLA - UCLA Newsroom

Anatomy

Sandra Oh Returns to TV as Series Regular … Just Not on ‘Grey’s Anatomy’ (Sorry, Fans) – Moviefone

Fun fact: Every five seconds, someone tweets "Come back to 'Grey's!" to "Grey's Anatomy" alum Sandra Oh. OK, maybe that's not a verified fact. It may even be too conservative a guess.

Oh recently teased fans by tweeting a photo with "Grey's" co-star Kevin McKidd (Dr. Owen Hunt), reviving the perpetual hope that Oh will bring Dr. Cristina Yang back to the ABC series. But before fans could get their hopes too high, Oh posted a link with news on her first post-"Grey's" series regular role -- leading the new BBC America series "Killing Eve."

Here's her "Grey's" reunion photo:

According to The Hollywood Reporter:

"Oh will take on the title role of Eve, a bored, whip-smart, pay-grade security services operative whose desk-bound job doesn't fulfill her fantasies of being a spy. Eve is pursued by Villanelle (uncast), an elegant, talented killer who clings to the luxuries her violent job affords her."

This marks Oh's first series regular role since leaving "Grey's Anatomy" after 10 seasons. She's had other roles -- recurring on "American Crime" and starring in movies like "Sideways" and the recent "Catfight" -- but this is her first full-time gig since playing Cristina Yang, and this time her character has her name in the title of the show. She's her own "person" now, Mer!

Fans replied to Oh's big news with ... dozens of requests for her to return to "Grey's." Sorry, but that's going to be her fate forever, or at least until she agrees to come back for the series finale or some other kind of closure. However, other fans are just happy to see more of this amazing actress on screen:

Want more stuff like this? Like us on Facebook.

More here:
Sandra Oh Returns to TV as Series Regular ... Just Not on 'Grey's Anatomy' (Sorry, Fans) - Moviefone

Human Behavior

This backflipping noodle has a lot to teach us about AI safety – The Verge

AI isnt going to be a threat to humanity because its evil or cruel, AI will be a threat to humanity because we havent properly explained what it is we want it to do. Consider the classic paperclip maximizer thought experiment, in which an all-powerful AI is told, simply, make paperclips. The AI, not constrained by any human morality or reason, does so, eventually transforming all resources on Earth into paperclips, and wiping out our species in the process. As with any relationship, when talking to our computers, communication is key.

Thats why a new piece of research published yesterday by Googles DeepMind and the Elon Musk-funded OpenAI institute is so interesting. It offers a simple way for humans to give feedback to AI systems crucially, without the instructor needing to know anything about programming or artificial intelligence.

The method is a variation of whats known as reinforcement learning or RL. With RL systems, a computer learns by trial-and-error, repeating the same task over and over, while programmers direct its actions by setting certain reward criteria. For example, if you want a computer to learn how to play Atari games (something DeepMind has done in the past) you might make the games point system the reward criteria. Over time, the algorithm will learn to play in a way that best accrues points, often leading to super-human performance.

What DeepMind and OpenAIs researchers have done is replace this predefined reward criteria with a much simpler feedback system. Humans are shown an AI performing two versions of the same task and simply tell it which is better. This happens again and again, and eventually the systems learns what is expected of it. Think of it like getting an eye test, when youre looking through different lenses, and being asked over and over: better... or worse? Heres what that looks like when teaching a computer to play the classic Atari game Q*bert:

This method of feedback is surprisingly effective, and researchers were able to use it to train an AI to play a number of Atari video games, as well perform simulated robot tasks (like picking telling an arm to pick up a ball). This better / worse reward function could even be used to program trickier behavior, like teaching a very basic virtual robot how to backflip. Thats how we get to the GIF at the top of the page. The behavior you see has been created by watching the Hopper bot jump up and down, and telling it well done when it gets a bit closer to doing a backflip. Over time, it learns how.

Of course, no one is suggesting this method is a cure-all for teaching AI. There are a number of big downsides and limitations in using this sort of feedback. The first being that although it doesnt take much skill on behalf of the human operator, it does take time. For example, in teaching the Hopper bot to backflip, a human was asked to judge its behavior some 900 times a process that took about an hour. The bot itself had to work through 70 hours of simulated training time, which was sped up artificially.

For some simple tasks, says Oxford Robotics researcher Markus Wulfmeier (who was not involved in this research), it would be quicker for a programmer to simply define what it is they wanted. But, says Wulfmeier, its increasingly important to render human supervision more effective for AI systems, and this paper represents a small step in the right direction.

DeepMind and OpenAI say pretty much the same its a small step, but a promising one, and in the future, theyre looking to apply it to more and more complex scenarios. Speaking to The Verge over email, DeepMind researcher Jan Leike said: The setup described in [our paper] already scales from robotic simulations to more complex Atari games, which suggests that the system will scale further. Leike suggests the next step is to test it in more varied 3D environments. You can read the full paper describing the work here.

See original here:
This backflipping noodle has a lot to teach us about AI safety - The Verge

Genetics

Discoveries in Genetics Are Changing the Way Drugs Are Tested – WCAI

Cystic fibrosis is a common genetic disease, relatively speaking. About one in 25 to 30 Caucasians are carriers of a cystic fibrosis mutation. But there are more than 1,700 mutations of the cystic fibrosis gene that can result in different disease symptoms.

Researchers and companies working on cystic fibrosis treatments are increasingly paying attention to which mutations a patient carries, and tailoring drugs to certain mutations.

Just recently, the FDA expanded its approval for one cystic fibrosis drug called Kalydeco. It had been approved for use in the case of ten of the mutations. Now its approved for 33.

They did it without a full clinical trial. The reason: each mutation is so rare, there just arent the hundreds of people a lab needs to do a clinical drug trial.

You just cant do it, said Art Caplan, Professor of Bioethics at New York University's Langone Medical Center.

Some of the things we usually see, like a randomized trial with hundreds or thousands of subjects, are not going to work because of the ability to pick out genetic differences among subjects, making it harder to do the big studies, he said.

Another fundamental shift in this area of medicine has to do with how drugs are used. Labs are now investigating how a the same drug might treat a wide range of maladies.

Some diseases that we dont even think of as related -- lets say, ALS, muscular dystrophy, and severe depression they may turn out to have the same chemical pathway that you can block with a drug, Caplan said.

It is huge and it is the future.

Originally posted here:
Discoveries in Genetics Are Changing the Way Drugs Are Tested - WCAI